Area
other (EverOS Cloud API — api.evermind.ai)
What happened?
Batch delete on the Cloud API consistently returns HTTP 500 SYSTEM_ERROR when called with user_id + session_id filters. Per the docs (llms-full.txt → "Delete Memories"), batch mode should accept {"user_id": ..., "session_id": ...} and return 204.
Single-mode delete ({"memory_id": "<memcell_id>"}) worked previously (last verified 2026-06-04), so this looks specific to batch mode — possibly a recent regression, or a 500 when the session filter matches no extracted memcells (either way, an unexpected 5xx rather than a documented 2xx/4xx).
Steps to reproduce
- Add messages for a session (response: 202
queued):
curl -X POST https://api.evermind.ai/api/v1/memories \
-H "Authorization: Bearer $EVEROS_API_KEY" -H "Content-Type: application/json" \
-d '{"user_id":"<uid>","session_id":"_schema_probe_video","messages":[{"role":"user","timestamp":1781138370000,"content":"probe"}]}'
-
(Optionally flush — returned {"status":"no_extraction"}.)
-
Batch delete the session:
curl -X POST https://api.evermind.ai/api/v1/memories/delete \
-H "Authorization: Bearer $EVEROS_API_KEY" -H "Content-Type: application/json" \
-d '{"user_id":"<uid>","session_id":"_schema_probe_video"}'
- Observe HTTP 500. Reproduced 3/3 times over ~30s, including after the flush in step 2.
Environment
EverOS Cloud, base URL https://api.evermind.ai (API v1), called via curl 2026-06-10 ~00:39–00:40 UTC.
Logs or screenshots
{"code":"SYSTEM_ERROR","message":"Internal server error","request_id":"-","timestamp":"2026-06-10T00:39:40.971744+00:00","path":"/api/v1/memories/delete"}
{"code":"SYSTEM_ERROR","message":"Internal server error","request_id":"-","timestamp":"2026-06-10T00:39:59.318123+00:00","path":"/api/v1/memories/delete"}
{"code":"SYSTEM_ERROR","message":"Internal server error","request_id":"-","timestamp":"2026-06-10T00:40:11.248256+00:00","path":"/api/v1/memories/delete"}
Note: request_id is "-" on these 500 responses, which also makes them hard to correlate on your side.
Area
other (EverOS Cloud API — api.evermind.ai)
What happened?
Batch delete on the Cloud API consistently returns HTTP 500
SYSTEM_ERRORwhen called withuser_id+session_idfilters. Per the docs (llms-full.txt → "Delete Memories"), batch mode should accept{"user_id": ..., "session_id": ...}and return 204.Single-mode delete (
{"memory_id": "<memcell_id>"}) worked previously (last verified 2026-06-04), so this looks specific to batch mode — possibly a recent regression, or a 500 when the session filter matches no extracted memcells (either way, an unexpected 5xx rather than a documented 2xx/4xx).Steps to reproduce
queued):(Optionally flush — returned
{"status":"no_extraction"}.)Batch delete the session:
Environment
EverOS Cloud, base URL https://api.evermind.ai (API v1), called via curl 2026-06-10 ~00:39–00:40 UTC.
Logs or screenshots
{"code":"SYSTEM_ERROR","message":"Internal server error","request_id":"-","timestamp":"2026-06-10T00:39:40.971744+00:00","path":"/api/v1/memories/delete"} {"code":"SYSTEM_ERROR","message":"Internal server error","request_id":"-","timestamp":"2026-06-10T00:39:59.318123+00:00","path":"/api/v1/memories/delete"} {"code":"SYSTEM_ERROR","message":"Internal server error","request_id":"-","timestamp":"2026-06-10T00:40:11.248256+00:00","path":"/api/v1/memories/delete"}Note:
request_idis"-"on these 500 responses, which also makes them hard to correlate on your side.